Skip to content

My implementation of Lagorce et al 2017: HOTS: A Hierarchy of Event-Based Time-Surfaces for Pattern Recognition

Notifications You must be signed in to change notification settings

Arata-Stu/event_based_time_surfaces

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

37 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

README

Overview

In order to familiarise myself with event-based cameras I implemented [1], in which "time surfaces" are generated in real-time from event camera feeds. Digits are classified by building time surfaces of each digit.

References

[1] Lagorce X, Orchard G, Galluppi F, Shi BE, Benosman RB. HOTS: A Hierarchy of Event-Based Time-Surfaces for Pattern Recognition. IEEE Trans Pattern Anal Mach Intell. 2017;39(7):1346‐1359. doi:10.1109/TPAMI.2016.2574707

Setup

Download the event-Python library from Github:

git clone https://github.com/Arata-Stu/event_based_time_surfaces.git

Install dependencies:

conda create -n <env_name>
conda activate env_name

Download the N-MNIST dataset from https://www.garrickorchard.com/datasets/n-mnist and place in ./datasets/mnist/

Train and visualise results

Run the Jupyter Notebook to train and visualise the digit classification:

jupyter notebook train_and_test_hots_model.ipynb

The training results for each of the layers will be shown. Here are screenshots from each of the layers:

About

My implementation of Lagorce et al 2017: HOTS: A Hierarchy of Event-Based Time-Surfaces for Pattern Recognition

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 73.1%
  • Jupyter Notebook 26.9%